Internetarchive

++++++++++++++++++

**Features and Improvements**

- Increased maximum identifier length from 80 to 100 characters in ``ia upload``.

**Bugfixes**

- As of version 2.11.0 of the requests library, all header values must be strings (i.e. not integers).
``internetarchive`` now converts all header values to strings.

++++++++++++++++++

**Features and Improvements**

- Added ``internetarchive.api.get_user_info()``.

++++++++++++++++++

**Bugfixes**

- Fixed bug where upload was failing on file-like objects (e.g. StringIO objects).

++++++++++++++++++

**Features and Improvements**

- All metadata writes are now submitted at -5 priority by default.
This is friendlier to the archive.org catalog, and should only be changed for one-off metadata writes.
- Expanded scope of valid identifiers in ``utils.validate_ia_identifier`` (i.e. ``ia upload``).
Periods are now allowed.
Periods, underscores, and dashes are not allowed as the first character.

++++++++++++++++++

**Features and Improvements**

- Search now uses the v1 scraping API endpoint.
- Moved ``internetarchive.item.Item.upload.iter_directory()`` to ``internetarchive.utils``.
- Added support for downloading "on-the-fly" files (e.g. EPUB, MOBI, and DAISY) via ``ia download <id> --on-the-fly`` or ``item.download(on_the_fly=True)``.

**Bugfixes**

- ``s3_is_overloaded()`` now returns ``True`` if the call is unsuccessful.
- Fixed bug in upload where a derive task wasn't being queued when a directory is uploaded.

++++++++++++++++++

**Features and Improvements**

- Use scrape API for getting total number of results rather than the advanced search API.
- Improved error messages for IA-S3 (upload) related errors.
- Added retry support to delete.
- ``ia delete`` no longer exits if a single request fails when deleting multiple files, but continues onto the next file.
If any file fails, the command will exit with a non-zero status code.
- All search requests now require authentication via IA-S3 keys.
You can run ``ia configure`` to generate a config file that will be used to authenticate all search requests automatically.
For more details refer to the following links:

http://internetarchive.readthedocs.io/en/latest/quickstart.html?highlight=configure#configuring

http://internetarchive.readthedocs.io/en/latest/api.html#configuration

- Added ability to specify your own filepath in ``ia configure`` and ``internetarchive.configure()``.

**Bugfixes**

- Updated ``requests`` lib version requirements.
This resolves issues with sending binary strings as bodies in Python 3.
- Improved support for Windows, see `https://github.com/jjjake/internetarchive/issues/126 <https://github.com/jjjake/internetarchive/issues/126>`_ for more details.
- Previously all requests were made in HTTP for Python versions < 2.7.9 due to the issues described at `https://urllib3.readthedocs.org/en/latest/security.html <https://urllib3.readthedocs.org/en/latest/security.html>`_.
In favor of security over convenience, all requests are now made via HTTPS regardless of Python version.
Refer to `http://internetarchive.readthedocs.org/en/latest/troubleshooting.html#https-issues <http://internetarchive.readthedocs.org/en/latest/troubleshooting.html#https-issues>`_ if you are experiencing issues.
- Fixed bug in ``ia`` CLI where ``--insecure`` was still making HTTPS requests when it should have been making HTTP requests.
- Fixed bug in ``ia delete`` where ``--all`` option wasn't working because it was using ``item.iter_files`` instead of ``item.get_files``.
- Fixed bug in ``ia upload`` where uploading files with unicode file names were failing.
- Fixed bug in upload where filenames with ``;`` characters were being truncated.
- Fixed bug in ``internetarchive.catalog`` where TypeError was being raised in Python 3 due to mixing bytes with strings.